crux-skill-memory-extract

To extract actionable learnings, analyze completed execution artifacts like logs and diffs to identify patterns, goals, and red flags, generating structured memory candidates for review.

Memory extraction in agent workflows preserves valuable project knowledge by analyzing execution artifacts and transforming discoveries into structured facts, preventing the loss of insights after work completion.

Prevent duplicate memory candidates by comparing proposed facts against existing memory collections to detect duplicates, identify related knowledge, and resolve conflicts before creating new memories.

Yes, you can generate ranked insights from completed specifications and task logs by reviewing execution states, classifying findings, and producing prioritized candidate memories with tags and evidence.

Yes, memory extraction works with post-execution workflows by validating completion states and analyzing code diffs, agent logs, and existing memory collections to produce structured candidate facts.

Automated knowledge management for completed tasks requires validation of completion state and produces candidate facts for user review, meaning it proposes ranked learnings rather than autonomously committing memories.